Medida de aceitação da corrupção: evidências de validade e confiabilidade do indicador

Abstract

The need to describe and establish relations of cause and effect of acceptance of corruption motivated researchers operationalize the concept. Measures of corrupt practices acceptance are often deployed in works that infers about cause and effect of the size of tolerance to the broader phenomenon of corruption. Four items of responses are the basis of the construct "Acceptance" and intended to measure the degree to which the respondent considers justifiable to practice acts as "tax evasion", "avoid public transportation tariff", "taking bribes" and "receiving government benefits without have direct them." Considering that constructs that claim to measure attitudes and values are subject to measurement errors, this work follows the methodological framework of Adcock and Collier (2001) and Carmines and Zaller (1979) and seeks evidence of validity and reliability of acceptance of corruption measure. In line with the findings of Catterberg et al (2013), this work argue that the attempt to measure the extent of acceptance may prevent bias caused by social desirability factor and the concept of corruption, as well as random errors caused by problems "non-action" and "non-response". The measurement error sources emerge through the background of the political culture, the dynamics of public opinion and at the time of collecting information for the survey methodology. The results suggest that despite the measure could grasp the underlying concept to be measured and differentiate from other constructs of other dimensions of the phenomenon of corruption, there is evidence that the latent variable has no significant internal consistency and the ability of observable indicators reflect the underlying dimension varies significantly over time.